Estimación del valor en riesgo (VaR) mediante simulaciones Montecarlo aplicado a un portafolio de acciones de renta variable de una entidad financiera pública.

Abstract

This research work aims to develop and implement a methodology for measuring market risk by estimating Value at Risk (VaR), based on a geometric Brownian motion for stocks and Montecarlo simulations. This applied to real data of a variable income portfolio of a public financial institution. The proposed methodology for the estimation of VaR is based on the modeling of the share price through a Geometric Brownian Movement, and the application of Monte Carlo simulations to establish the possible trajectories for the share price. With all this, the application with real data of the stock portfolio of the financial institution whose portfolio is made up of 19 assets. And, in addition, the historical database includes the period from July 2020 to December 2021; this results in adequate validation measures (MAPE and RMSE). And a p-value to the Kupiec and Christoffersen hypothesis tests that suggest accepting the proposed hypotheses, being that the VaR estimate with a cutoff date of January 3, 2022, is appropriately measuring the level of risk.